Risk factors and outcomes of metastatic poorly differentiated thyroid carcinoma

Abstract

BACKGROUND

Poorly differentiated thyroid carcinoma (PDTC) is a rare and aggressive form of thyroid cancer. Distant metastasis occurs frequently in PDTC.

AIM

To determine factors associated with distant metastasis and the effects of metastasis, either diagnosed on initial presentation or developing during follow-up, on mortality in PDTC patients.

METHODS

Patients with PDTC diagnosed between January 1, 1985 and July 31, 2022 were identified using a thyroid cancer database at a medical center in Taiwan. Factors associated with distant metastasis and cancer-specific survival (CSS) were analyzed using binary logistic analysis and Cox regression, respectively. Survival analysis was conducted using the Kaplan–Meier method.

RESULTS

The study cohort included 39 patients with PDTC, including 16 with distant metastasis on initial presentation, 5 with metastasis during the follow-up period, and 18 with no evidence of metastasis. Older age (≥ 45 years) was significantly associated with a higher risk of distant metastasis (odds ratio: 5.31; 95% confidence interval: 1.27–22.2; P = 0.018), while sex, tumor size, T stage, and N stage were not associated with distant metastasis. Patients with metastatic PDTC, either diagnosed at presentation or developing during follow-up, had poorer 5-year CSS rates (55.0% and 40.0%, respectively) than those without metastasis (5-year CSS, 93.8%) (P = 0.001 for both comparisons).

CONCLUSION

Older patients with PDTC have an increased risk of distant metastasis. Patients with metastatic PDTC, both diagnosed at presentation and developing during follow-up, have a dismal prognosis.

Key Words: Poorly differentiated thyroid carcinoma; Age; Lymph node metastasis; Distant metastasis; Survival

Core Tip: Poorly differentiated thyroid carcinoma (PDTC) is a rare but aggressive subtype of thyroid cancer. A significant proportion of patients with PDTC have distant metastasis. Patients with metastatic PDTC diagnosed at presentation or developing during follow-up have dismal 5-year cancer-specific survival rates of only 55.0% and 40.0%, respectively. Older age (≥ 45 years) is a risk factor for metastatic PDTC.

INTRODUCTION

The incidence of thyroid cancer has increased over the past three decades[1]. The majority of thyroid cancers are differentiated thyroid carcinomas (DTC), including papillary thyroid carcinoma, follicular thyroid carcinoma, and oncocytic cell carcinoma. Most DTC patients have a favorable prognosis, with a 10-year survival rate of approximately 90%. Anaplastic thyroid carcinoma (ATC) is an aggressive and fatal disease, with a 2-year survival rate of only 10%–15%. Poorly differentiated thyroid carcinoma (PDTC) ranks between DTC and ATC with respect to the degree of differentiation. PDTC represents 3–5% of all thyroid cancer cases, with a 5-year survival rate of 47.0%–89.3%[2-4]. Multiple prognostic factors associated with poor survival in PDTC patients have been reported, including older age (≥ 45 years), larger tumor size (> 4 cm), extrathyroidal extension, and distant metastasis at presentation[5,6].

Distant metastases are observed at presentation in 26.0%–65.5% of PDTC patients. The lung and bone are the most common metastatic sites[5-7]. Among these patients, the 5-year cancer-specific survival (CSS) is only 34%[5]. A significant proportion of PDTC patients (22%–50%) develop metastatic disease during follow-up[6,8]. However, few studies have addressed the clinical course of patients who develop metastatic PDTC during follow-up.

This study aimed to identify factors associated with PDTC metastasis and to assess the effect of distant metastasis diagnosed upon initial presentation or during follow-up on CSS.

MATERIALS AND METHODS

Source of data

This study examined data from a thyroid cancer database at Chang Gung Memorial Hospital in Taiwan. The database was specifically developed to support thyroid cancer research[9]. Data were collected prospectively for all patients receiving thyroid cancer treatment at this medical center. Comprehensive records, including clinical characteristics (age, sex), laboratory results (thyroid function tests, thyroglobulin, anti-thyroglobulin antibody), imaging data [chest radiography, computed tomography (CT), magnetic resonance imaging, radioactive iodine (RAI) scan, bone scan], pathological findings, treatment regimens [surgery, RAI therapy, external beam radiation therapy (EBRT), chemotherapy, targeted therapy, immunotherapy], clinical course, and outcomes were collected and maintained.

Study cohort

A total of 47 patients with PDTC diagnosed between January 1, 1985 and July 31, 2022 were identified in the database. The exclusion criteria included having primary surgery at another hospital, loss to follow-up, or missing data. The eligible patients were assigned to one of three groups according to the timing of distant metastasis diagnosis, as follows: M0, no evidence of metastasis during the study period; M1, metastasis identified on initial presentation; and M2, appearance of metastasis during the follow-up period.

Covariates

Covariates examined in survival analysis included age, sex, tumor size, T stage, N stage, extrathyroidal extension, type of surgery, RAI treatment, and EBRT[3-6]. Tumor staging was performed in accordance with the eighth edition of the AJCC staging system[10].

Treatment protocol and follow-up

Thyroidectomy was the first-line treatment for PDTC in this study. Neck lymph nodes were removed by neck dissection if metastasis was suspected. RAI was given if deemed appropriate by the treating physician. Following primary treatment, levothyroxine was prescribed for thyroid hormone supplementation or thyroid-stimulating hormone suppression. Assessments of serum thyroglobulin and antithyroglobulin antibody and cervical ultrasound were typically performed every 3-12 months during the follow-up period. Imaging assessments including chest X-ray, RAI uptake and scan, CT, and ¹⁸F-fluorodeoxyglucose positron emission tomography/CT were performed when metastasis was suspected. EBRT was used for locoregional control and palliation of symptomatic metastasis. Four patients with metastatic PDTC were treated with targeted therapy including lenvatinib, sorafenib, and cabozantinib.

Statistical analysis

Continuous variables are presented as the median [interquartile range (IQR)]. Categorical variables are presented as a number (percentage). The Kruskal-Wallis H test was used to analyze continuous variables not following a normal distribution, and the χ² test was used to analyze categorical variables. Cox regression analysis was used to determine the hazard ratio (HR) of risk factors associated with CSS. Binary logistic regression was used to calculate the odds ratio (OR) of risk factors for distant metastasis. In multivariable analysis, only the independent variables were included for further evaluation. Survival analysis was conducted using the Kaplan–Meier method. Statistical analyses were performed using SPSS statistical software (version 22.0, SPSS Inc., IBM corp., Armonk, NY, United States). P < 0.05 was considered statistically significant.

Ethics statement

This study was approved by the Chang Gung Medical Foundation Institutional Review Board (No. 202401356B0). The Institutional Review Board waived the requirement for obtaining informed consent. The study was performed in accordance with the principles of the Declaration of Helsinki.

RESULTS

Clinical characteristics

A total of 47 patients with PDTC were identified in our thyroid cancer database. Eight patients were excluded, including two who underwent primary cancer surgery at other hospitals, two who were lost to follow-up, and four who had missing data. Thus, 39 PDTC patients were included in the final analysis. These patients were divided into three groups based on the presence of metastasis or not: M0, no evidence of metastasis (n = 18); M1, metastasis on initial presentation (n = 16); and M2, metastasis that developed during follow-up (n = 5) (Figure 1).

Figure 1 Flowchart of cohort establishment.

The clinical characteristics of the study cohort are shown in Table 1. The median age at PTDC diagnosis was highest in the M2 group (58 years), followed by the M1 group (52 years) and the M0 group (41 years) (P = 0.042). The highest percentage of female patients was discovered in the M2 group (100%), followed by the M0 group (72.2%), and the M1 group (43.8%) (P = 0.045). Differences among the three cohorts in the proportion of older patients (≥ 45 years), tumor size (≤ 2 cm, > 2–4 cm, and > 4 cm), T stage, N stage, and extrathyroidal invasion did not reach statistical significance (P > 0.05). The most common site of metastasis was the lung in the M1 (56.3%) and M2 (60.0%) groups. The other metastatic organs included bone, the lung and bone, the lung and brain, and bone and the brain. The M0 group and M2 group had the largest percentage of patients who had total thyroidectomy (100%, both groups), compared to the M1 group (75%) (P = 0.041). The proportions of patients who received RAI were similar in the M0 group (61.1%), M1 group (75.0%), and M2 group (100%) (P = 0.217). The median cumulative dose of RAI was highest in the M2 group (580 mCi), followed by the M1 group (100 mCi) and the M0 group (75 mCi) (P = 0.005). The percentage of patients who had EBRT was highest in the M2 group (60.0%), followed by the M1 group (43.8%) and the M0 group (16.7%). However, these differences among the three groups were not statistically significant (P = 0.099). The median dose of EBRT was similar in all the three groups (M0 group, 5600 cGy; M1 group, 3720 cGy; and M2 group, 5000 cGy; P = 0.634). Targeted therapy was used in each group as follows: M1 group, 3 patients; M2 group, 1 patient; and M0 group, no patients. The M2 group had the highest cancer-specific mortality (100%), followed by the M1 group (62.5%) and the M0 group (5.6%) (P = 0.001). The M2 group had the worst overall mortality (100%), followed by the M1 group (75.0%) and the M0 group (11.1%) (P = 0.001). The median follow-up duration in the M0, M1, and M2 groups was 8.0 years, 2.0 years, and 4.2 years, respectively (P = 0.156).

Table 1 Characteristics of 39 poorly differentiated thyroid carcinoma patients, n (%).

	M0 (n = 18)	M1 (n = 16)	M2 (n = 5)	P value
Age, median (years, IQR)	41 (26-59)	52 (49-63)	58 (51-62)	0.042
Age				0.060
< 45 years	10 (55.6)	3 (18.8)	1 (20.0)
≥ 45 years	8 (44.4)	13 (81.2)	4 (80.0)
Sex				0.045
Female	13 (72.2)	7 (43.8)	5 (100)
Male	5 (27.8)	9 (56.2)	0
Tumor size (cm)				0.222
≤ 2	3 (16.7)	0	0
> 2-4	6 (33.3)	6 (37.5)	3 (75.0)
> 4	9 (50.0)	10 (62.5)	1 (25.0)
T stage				0.536
T1	3 (16.7)	0	0
T2	4 (22.2)	5 (31.2)	1 (20.0)
T3	5 (27.8)	3 (18.8)	1 (20.0)
T4	6 (33.3)	8 (50.0)	3 (60.0)
N stage				0.323
N0	17 (94.4)	13 (81.2)	5 (100)
N1	4 (10.3)	3 (18.8)	0
Extrathyroidal invasion	6 (33.3)	8 (50.0)	4 (80.0)	0.166
Initial metastatic sites
Lung	0	9 (56.3)	3 (60.0)
Bone	0	3 (18.8)	0
Lung, bone	0	2 (12.5)	2 (40.0)
Lung, brain	0	1 (6.3)	0
Bone, brain	0	1 (6.3)	0
Type of thyroid surgery				0.041
Lobectomy or subtotal thyroidectomy	0	4 (25.0)	0
Total thyroidectomy	18 (100)	12 (75.0)	5 (100)
131I therapy	11 (61.1)	12 (75.0)	5 (100)	0.217
131I cumulative dose, median (mCi, IQR)	75 (15-100)	100 (97-162)	580 (100-835)	0.005
External beam radiation therapy	3 (16.7)	7 (43.8)	3 (60.0)	0.099
External beam radiation therapy, median (cGy, IQR)	5600 (3950-6200)	3720 (1725-5310)	5000 (4750-5500)	0.634
Targeted therapy	0	3 (18.8)	1 (20.0)	-
Lenvatinib	0	1 (6.25)	0
Lenvatinib, sorafenib	0	2 (12.5)	0
Lenvatinib, sorafenib, cabozantinib	0	0	1 (20.0)
Cancer-specific mortality	1 (5.6)	10 (62.5)	5 (100)	0.001
Overall mortality	2 (11.1)	12 (75.0)	5 (100)	0.001
Follow-up, median (years, IQR)	8.0 (2.4-14.3)	2.0 (1.4-5.2)	4.2 (2.8-8.1)	0.156

IQR: Interquartile range.

Factors associated with cancer-specific mortality

We sought to identify factors associated with cancer-specific mortality in PDTC patients (Table 2). Univariate Cox regression analysis showed that cancer-specific mortality was associated with N stage [HR: 6.86; 95% confidence interval (CI): 1.82–25.8; P = 0.004], M stage (HR: 2.96; 95%CI: 1.53–5.74; P = 0.001), and EBRT (HR: 4.42; 95%CI: 1.50–13.0; P = 0.007). Using multivariate Cox regression analysis adjusted for N stage, M stage, and EBRT, we found that N1 stage (adjusted HR: 10.1; 95%CI: 2.34–43.3; P = 0.002) and distant metastasis (adjusted HR: 2.41; 95%CI: 1.07–5.39; P = 0.033) were significantly associated with cancer-specific mortality.

Table 2 Risk factors associated with cancer-specific mortality in poorly differentiated thyroid carcinoma patients, n (%).

	Survival (n = 23)	Mortality (n = 16)	Univariate HR (95%CI)	P value	Multivariate HR (95%CI)1	P value
Age
< 45 years	11 (47.8)	3 (18.8)	Reference
≥ 45 years	12 (52.2)	13 (81.2)	3.47 (0.95–12.6)	0.060
Sex
Female	14 (60.9)	11 (68.8)	Reference
Male	9 (39.1)	5 (31.2)	1.58 (0.52–4.81)	0.416
Tumor size
≤ 2 cm	3 (13.0)	0	Reference
> 2-4 cm	7 (30.4)	8 (53.3)
> 4 cm	13 (56.5)	7 (46.7)	1.39 (0.62–3.12)	0.423
T stage
T1/T2	8 (34.8)	5 (31.2)	Reference
T3/T4	15 (65.2)	11 (68.8)	1.31 (0.44–3.85)	0.627
N stage
N0	23 (100)	12 (75.0)	Reference		Reference
N1	0	4 (25.0)	6.86 (1.82–25.8)	0.004	10.1 (2.34–43.3)	0.002
M stage
M0	17 (73.9)	1 (37.5)	Reference		Reference
M1	6 (26.1)	10 (62.5)
M2	0	5 (31.2)	2.96 (1.53–5.74)	0.001	2.41 (1.07–5.39)	0.033
Extrathyroidal invasion
No	14 (60.9)	7 (43.8)	Reference
Yes	9 (39.1)	9 (56.2)	1.64 (0.61–4.43)	0.329
Type of thyroid surgery
Total thyroidectomy	21 (91.3)	14 (87.5)	Reference
Lobectomy or subtotal thyroidectomy	2 (4.3)	2 (6.2)	3.48 (0.71–17.0)	0.124
131I cumulative dose
< 100 mCi	14 (60.9)	7 (43.8)	Reference
≥ 100 mCi	9 (39.1)	9 (56.2)	1.42 (0.52–3.83)	0.493
External beam radiation therapy
No	20 (87.0)	6 (37.5)	Reference		Reference
Yes	3 (13.0)	10 (62.5)	4.42 (1.50–13.0)	0.007	3.07 (0.86–11.0)	0.084

¹Adjusted for N stage, M stage, and external beam radiation therapy.

Clinical features of five patients who developed metastasis during follow-up

Five patients developed distant metastasis during follow-up. The clinical features of these patients are shown in Table 3. All of these patients were female, aged 44–77 years. Most patients (60.0%) had T4 stage, followed by T3 stage (20.0%) and T2 stage (20.0%). No patients had metastatic neck lymph nodes at presentation. All patients (100%) received total thyroidectomy, and four patients (80%) underwent RAI treatment with a cumulative dose of 30–160 mCi before metastasis was discovered. The time interval between the discovery of distant metastasis and the diagnosis of PDTC ranged from 0.7 to 11.9 years. More distant metastases were identified by chest X-ray (40%), followed by RAI scan (20%), CT (20%), and positron emission tomography/CT (20%). The sites of metastasis included the lung (60%) and both the lung and bone (40%). Three patients (60%) received palliative EBRT (range, 4500–6000 cGy). One patient (20%) received targeted therapy (sequential lenvatinib, sorafenib, and cabozantinib treatment). All patients had RAI treatment for PDTC, including three patients (60%) with a cumulative RAI dose ≥ 580 mCi. All patients died of PDTC, with the time from the diagnosis of metastasis to death within 5.9 years (range, 0.6–5.9 years).

Table 3 Clinical features of five patients who developed metastatic disease during follow-up.

Patient number	Age/Sex	TNM at diagnosis	Type of thyroid surgery	Time from PDTC diagnosis to metastasis (years)	Cumulative dose of RAI before metastasis (mCi)	Positive finding at detection	Sites of metastasis	Time from metastasis to death (years)	Cumulative dose of RAI (mCi)	External beam radiation therapy, dose (cGy)	Targeted therapy
1	44/F	T4N0M0	Total thyroidectomy	11.9	0	CT	Lung, bone	5.9	980	6000
2	51/F	T4N0M0	Total thyroidectomy	1.7	100	PET/CT	Lung	2.5	100	0	Lenvatinib, sorafenib, cabozantinib
3	58/F	T4N0M0	Total thyroidectomy	0.7	30	RAI scan	Lung	2.1	580	5000
4	62/F	T2N0M0	Total thyroidectomy	4.9	160	Chest X-ray	Lung, bone	3.2	835	4500
5	77/F	T3N0M0	Total thyroidectomy	0.9	30	Chest X-ray	Lung	0.6	60	0

CT: Computed tomography; PET: Positron emission tomography; RAI: Radioactive iodine; PDTC: Poorly differentiated thyroid carcinoma.

Factors associated with distant metastasis

In these 39 PDTC patients, 16 (41.0%) had metastasis on initial presentation and 5 (12.8%) had metastasis during follow-up. Factors associated with metastasis were assessed (Table 4). Our data demonstrate that distant metastasis was significantly associated with older age (≥ 45 years) (OR: 5.31; 95%CI: 1.27-22.2; P = 0.018), while sex, tumor size, T stage, and N stage did not correlate with distant metastasis. The lung was the most common metastatic site for PDTC, a finding consistent with a previous study[5].

Table 4 Risk factors associated with distant metastasis in poorly differentiated thyroid carcinoma patients, n (%).

	M0 (n = 18)	M1 + M2 (n = 21)	Univariate OR (95%CI)	P value
Age
< 45 years	10 (55.6)	4 ((19.0)	Reference
≥ 45 years	8 (44.4)	17 (81.0)	5.31 (1.27–22.2)	0.018
Sex
Female	13 (72.2)	12 (57.1)	Reference
Male	5 (27.8)	9 (42.9)	1.95 (0.51–7.49)	0.328
Tumor size
≤ 2 cm	3 (16.7)	0	Reference
> 2-4 cm	6 (33.3)	9 (45.0)
> 4 cm	9 (50.0)	11 (55.0)	1.72 (0.62–4.83)	0.301
T stage
T1/T2	7 (38.9)	6 (28.6)	Reference
T3/T4	11 (61.1)	15 (71.4)	1.59 (0.42–6.07)	0.496
N stage
N0	17 (94.4)	18 (85.7)	Reference
N1	1 (5.6)	3 (14.3)	2.83 (0.27–29.9)	0.370

Cancer-specific survival

Our results show that patients who died of PDTC were more likely to have lymph node metastasis and distant metastasis. Kaplan–Meier survival analysis stratified by N stage and M stage showed that the N1 group had poorer 5-year and 10-year CSS (25.0% and 0%) than the N0 group (73.3% and 63.1%) (P = 0.001) (Figure 2A). Patients with M1 and M2 disease were more likely to die of the disease than those with M0 disease (Figure 2B). Poorer 5-year and 10-year CSS rates were seen in the M1 group (55.0% and 29.3%) and M2 group (40.0 and 20.0%), compared with the M0 group (93.8% and 93.8%) (P = 0.001 for both comparisons). CSS was similar between the M1 group and M2 group (P = 0.831).

Figure 2 Kaplan–Meier curves of cancer-specific survival in poorly differentiated thyroid carcinoma patients. A: Kaplan–Meier survival curves for patients with (N1) and without (N0) lymph node metastasis; B: Kaplan–Meier survival curves for patients with distant metastasis at presentation (M1), distant metastasis developing during follow-up (M2), and no metastasis (M0); C: Kaplan–Meier survival curves for poorly differentiated thyroid carcinoma patients with stages I, II, III, and IV disease at diagnosis.

The 8^th edition of the AJCC provides the same definition of TNM for PDTC and DTC[10]. This staging system includes a stage classification to stratify the risk of mortality for patients with DTC, but that classification is not indicated for patients with PDTC. We sought to analyze the CSS of these 39 PDTC patients based on the 8^th edition of the stage classification in DTC patients aged ≥ 55 years. Our data reveal that PDTC patients with stage IV disease had worse 5-year and 10-year CSS (54.6% and 34.1%) than those with stage I disease (100% and 83.3%) (P = 0.029). However, the differences in survival between the other cohorts did not reach statistical significance (P > 0.05 for 4 comparisons) (Figure 2C).

DISCUSSION

We found that PDTC patients without distant metastasis had a promising outcome, with a 5-year CSS rate of 93.8%. In contrast, the prognosis for patients with metastatic PDTC, whether diagnosed upon presentation or during follow-up, was dismal, with 5-year CSS rates of 55% and 40%, respectively. Our data are consistent with a previous report showing a 5-year CSS rate of 34% for patients with metastatic PDTC diagnosed at presentation[5]. We additionally showed that PDTC patients who developed metastasis during follow-up had similarly dismal outcomes to those who had metastasis at diagnosis.

Our study revealed that a substantial percentage of PDTC patients with M0 disease (21.7%) initially eventually developed distant metastasis. The time interval between the diagnosis of PDTC and the identification of metastasis ranged from 0.7 to 11.9 years. Most metastases (80%) occurred within 5 years after the diagnosis of PDTC in older patients (≥ 51 years), suggesting that older patients should be monitored more intensively within 5 years. The majority of patients (80%) had advanced T stage (≥ T3). However, one patient (20%) who had T2N0M0 disease initially developed lung and bone metastases 4.9 years after PDTC diagnosis and died subsequently. These data are consistent with a previous report showing that a small proportion of patients with T1/T2 disease (6%) would succumb to PDTC[5]. The most common metastatic sites in the M2 group were the lung (100.0%) and bone (40.0%), which was similar to the M1 group (lung, 75.0%; bone, 31.2%).

Previous reports have demonstrated that older age (> 45 years), T4a stage, extrathyroidal extension, tumor mitosis and necrosis, and distant metastasis at presentation were predictors of cancer-specific mortality in PDTC patients[3,5,6,7,11]. In this study, we found that metastatic PDTC, either cervical lymph node metastasis or distant metastasis, was an independent risk factor correlating with cancer-specific mortality, while age, sex, tumor size, T stage, extrathyroidal invasion, type of thyroid surgery, RAI cumulative dose, and EBRT were not. Our data indicate that metastasis, whether to lymph nodes or solid organs, is the most important prognostic factor for PDTC mortality.

In this study, a total of 53.8% patients had metastatic PDTC. The risk factor correlated with metastasis was older age. Understanding the risk factors for distant metastasis might allow clinicians to plan more intensive surveillance of patients who are at risk of metastasis. A recent study revealed that patient age was significantly associated with distant metastasis in papillary thyroid microcarcinoma patients[12].

A previous study has shown that aggressive surgery including total thyroidectomy and compartment neck dissection in PDTC patients with gross extrathyroidal extension can yield promising locoregional control, with a 5-year local recurrence–free survival and regional recurrence–free survival of 70% and 62%, respectively[6]. In this study, patients in the M2 group had undergone total thyroidectomy as the primary surgery, but all of them developed metastasis and died of PDTC eventually. These results reflect the aggressive and fatal nature of PDTC.

The majority of our patients (71.8%) received RAI therapy. We found that a higher dose of RAI (≥ 100 mCi) was not associated with better survival. A recent study showed that RAI did not improve CSS in PDTC patients[13]. The reason for the ineffectiveness of RAI in this and previous studies is unclear. One study reports that most patients with metastatic PDTC (70%) were RAI-avid[6]. The RAI uptake in metastatic PDTC may not exceed the therapeutic threshold, likely due to varying mixtures of well- and less-well–differentiated tumor components, which may account for the limited efficacy of RAI in PDTC[6,7]. In addition, resistance to RAI might be another cause for the limited efficacy of RAI for treating PDTC. Currently, RAI is recommended as an adjuvant therapy for PDTC patients with distant metastasis that is RAI-avid[6,14,15].

One-third of our patients received EBRT. We found that EBRT did not improve CSS, which is consistent with previous studies reporting that EBRT did not improve survival in PDTC patients[16,17]. EBRT is mainly used for local tumor control and palliative treatment of metastatic disease[16].

Lenvatinib, sorafenib, and cabozantinib are tyrosine kinase inhibitors that have been approved for the treatment of progressive radioiodine-refractory DTC based on the results of the SELECT, DECISION, and COSMIC-311 studies, respectively[18-20]. Although the SELECT study aimed to evaluate the effects of lenvatinib in patients with DTC, that study included 47 patients with PDTC, including 28 patients in the lenvatinib group and 19 patients in the control group. The results show that lenvatinib improved progression-free survival over that of control treatment in PDTC patients (HR: 0.21; 95%CI: 0.08–0.56)[18]. In our study, four PDTC patients received a tyrosine kinase inhibitor: Lenvatinib, n = 1; lenvatinib initially followed by sorafenib, n = 1; sorafenib initially followed by lenvatinib, n = 1; and sequential lenvatinib, sorafenib, and cabozantinib, n = 1. Disease progression was observed in the majority of the patients who underwent tyrosine kinase inhibitor treatment (Supplementary Figure 1). Novel therapies are needed to increase outcomes for patients with advanced PDTC.

Some limitations of this study deserve mentioning. First, the utility of the 8^th edition of TNM staging system to stratify the risk of mortality for PDTC patients is unclear and warrants further studies with a larger sample size. Second, we sought to identify factors associated with lymph node metastasis. The results show that age, sex, tumor size, T stage, M stage, and extrathyroidal invasion did not correlate with lymph node metastasis (P > 0.05) (Supplementary Table 1). Third, we did not assess genetic alterations in these PDTC patients. High frequencies of genetic mutations have been found in PDTC, including RAS, BRAF, TERT promoter, TP53, and ATM[21-27]. A previous study showed that RAS-mutated PDTC had an elevated rate of distant metastasis, and BRAF-mutated PDTC had an elevated rate of lymph node metastasis[28]. TERT promoter mutations are usually associated with an aggressive phenotype and poor outcomes[24]. Previous studies have revealed that patients with fatal PDTC had a higher frequency of mutations in HRAS, BRAF, TERT promoter, TP53 and ATM than those with nonfatal PDTC[24,29]. Genetic testing should be performed for patients with refractory PDTC to identify potential cancer-related molecular targets, particularly since the treatment of thyroid cancer has undergone major advancements in recent years[30-34]. Selpercatinib and pralsetinib have been approved for advanced thyroid cancer with RET alterations[30,31]. Larotrectinib and entrectinib were approved for advanced thyroid cancer with NTRK fusions[32,33]. Combination treatment with dabrafenib and trametinib is indicated for solid tumors with the BRAF^V600E mutation[34]. Fourth, this study included only 39 PDTC patients. The small sample size may limit the statistical power of this study. Additional studies with larger cohorts will be needed to confirm our results. Fifth, the heterogeneity in treatment approaches, including variations in surgery, RAI therapy, EBRT, and targeted therapy, may influence survival outcomes in this study.

The incidence of PDTC was 0.69% of all thyroid cancer patients in our dataset between 1985 and 2022. In Taiwan, the incidence of PDTC was < 0.48% of all thyroid cancer cases from 2001 to 2022[35]. These data indicate that PDTC is a rare disease in Taiwan. The risk factors associated with PDTC in Taiwan are unclear and need further investigation.

CONCLUSION

We observed that up to 41.0% of patients with PDTC presented with distant metastasis. In addition, 12.8% developed distant metastasis during follow-up. The survival of patients with distant metastasis, either diagnosed at presentation or during follow-up, is poor. Older age is associated with a higher risk of metastatic PDTC.

ACKNOWLEDGEMENTS

We would like to thank the Center for Big Data Analytics and Statistics at Chang Gung Memorial Hospital at Linkou, Taiwan for their assistance.